ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, May 1976, Copyright 0 1976 American Society for Microbiology
p.
766-770
Vol. 9, No. 5 Printed in U.SA.
Active Immobilized Antibiotics Based on Metal Hydroxides1 JOHN F. KENNEDY* AND JOHN D. HUMPHREYS
Department of Chemistry, University of Birmingham, Edgbaston, Birmingham B15 2TT, England Received for publication 22 December 1975
The water-insoluble hydroxides of zirconium (IV), titanium (IV), titanium (III), iron (II), vanadium (Ill), and tin (II) have been used to prepare insoluble derivatives of a cyclic peptide antibiotic by a facile chelation process. Testing of the antibacterial activities of the products against two gram-positive and two gram-negative bacteria showed that in the majority of cases the water-insoluble antibiotics remained active against those bacteria susceptible to the parent antibiotic. The power of the assay system has been extended by the novel use of colored organisms to aid determinations where the growth of normal organisms could not be distinguished from the appearance of the supporting material. Insoluble derivatives of neomycin, polymyxin B, streptomycin, ampicillin, penicillin G, and chloramphenicol were prepared by chelation with zirconium hydroxide, and these derivatives similarly reflected the antibacterial activities of the parent compounds. Several of the metal hydroxides themselves possess antibacterial activity due to complex formation with the bacteria. However, the use of selected metal hydroxides can afford a simple, inexpensive, and inert matrix for antibiotic immobilization, resulting in an antibacterial product that may possess slow-release properties. The mechanisms by which the metal hydroxide-antibiotic association-dissociation may occur are discussed.
The potential uses of active, insolubilized antibiotic derivatives for the protection of biodegradable substances from microbial attack and for slow-release formulations in medical and industrial fields have already been discussed (4-6). Matrixes used for the formation of such derivatives have included cellulose (5), cellulose carbonate (5), poly(N-acryloyl-4- and -5aminosalicylic acids) (6) and cellulose-metal chelates (4). Whereas the formation of a completely covalent bond between the antibiotic and the chosen insoluble matrix provides the most stable immobilized antibiotic system, such a reaction may not always be feasible in practice and complete stability may not be desirable (e.g., in slow-release situations). Thus, the use of a simpler means of attachment can confer advantages of ease of preparation and also of economy. This has been demonstrated by the use of cellulose-metal chelates to insolubilize a range of antibiotics with retention of antibacterial activity (4). However, even in these cases, prior preparation of the activated matrix for immobilization is necessary. We now report the immobilization of antibiotics by a simple coupling process to matrixes which may be prepared as required without any need for specialized equipment. The matrixes are the hydroxides (hydrous oxides, hydroxyoxides) of the metal components of the cellulose-metalI
U.S. Patent 3,912,593.
antibiotic chelates (4). Such matrixes have already been used for the insolubilization of enzymes (1, 7), and the extension ofthis technique to the formation of insolubilized antibiotics has been studied and is now reported. MATERIALS AND METHODS Antibiotics. Antibiotics were obtained as follows: lathumycin, Gist-Brocades N.V.; neomycin sulfate, Boots Pure Drug Co. Ltd.; streptomycin sulfate, Glaxo Laboratories Ltd.; polymyxin B sulfate, Burroughs-Wellcome & Co.; ampicillin, Beecham Research Laboratories; chloramphenicol, Aspro-Nicholas Ltd.; penicillin G sodium salt, Imperial Chemical Industries, Ltd., Pharmaceuticals Division. Bacteria. The sources of bacteria used were: Escherichia coli NCTC 86, Pseudomonas aeruginosa BUCD 53 (Department of Chemistry, University of Birmingham), Streptococcus faecalis NCTC 370, Staphylococcus aureus (formerly S. pyogenes) NCTC 7447, Serratis marcescens BUCD 30, and Chromobacterium violaceum NCIB 9131. Metal chlorides. Iron (III) chloride (Hopkin and Williams Ltd.), tin (II) chloride (BDH Chemicals Ltd.), vanadium (III) chloride (BDH Chemicals Ltd.), and zirconium (IV) chloride (BDH Chemicals Ltd.) were standard laboratory reagents. Titanium (III) chloride (BDH Chemicals Ltd.) was supplied as a 12.5% (wt/vol) solution in hydrochloric acid and titanium (IV) chloride (BDH Chemicals Ltd.) was supplied as a 15% (wt/vol) solution in hydrochloric acid (5 N). Preparation of metal hydroxides. A standard method was used for the preparation of the various 766
VOL. 9, 1976
METAL HYDROXIDES AND IMMOBILIZED ANTIBIOTICS
metal hydroxides from solutions of the appropriate metal chloride above. The solutions of the metal chlorides were prepared by dissolving the requisite amount of the metal chloride in hydrochloric acid (1.0 M), except for the solution of tin (II) chloride, which was prepared by dissolving the chloride in 5.0 M hydrochloric acid to give final concentrations of 0.65 M. Titanium chloride solutions were used as supplied. The metal hydroxides were precipitated by the slow addition of ammonium hydroxide (2 M) to the metal chloride solutions (1.3 mmol of metal) to give the desired pH and to ensure that the gelatinous suspension formed was homogenous. Coupling of antibiotics to metal hydroxides. Two suspensions of the hydroxides of each of the metals listed above were prepared as described above, one suspension being adjusted to pH 5 and the other to pH 7. To every sample was added a 0. 1% (wt/vol) solution (5.0 ml) of lathumycin, a cyclic peptide antibiotic, and distilled water to give a final volume of 10.0 ml. After stirring for 2 h at 22 C, the mixtures were centrifuged and the absorbances of the supernatants were measured (to determine the amount of antibiotic remaining in solution). After washing with water (3 x 5.0 ml) the samples were tested for antibacterial activity as described below. A further six antibiotics were coupled to zirconium hydroxide at pH 7 in a similar manner to that described above, and the metal hydroxides (pH 7.0) were also prepared without any added antibiotic, to act as controls. Antibacterial testing of immobilized antibiotics. Ditch plates were prepared by allowing nutrient agar to solidify in a petri dish and a strip of agar, approximately 1 cm wide, was removed to form the ditch. The samples of immobilized antibiotics were resuspended in water (total volume, 10 ml), and a portion of the suspension was mixed with an equal volume of double-strength agar (1.6%, wt/vol) and introduced into the ditch. The plates were inoculated (in stripes perpendicular to the ditch) with two gram-negative (E. coli and P. aeruginosa) and two gram-positive (S. faecalis and S. aureus) bacteria, four stripes in all. In one case, two additional gramnegative bacteria were used (S. marcescens and C. violaceum) for which a second plate was used, two stripes in all.
767
The plates were incubated at 37 C (25 C for S. and C. violaceum) for 18 to 24 h, and then the degree of inhibition of growth of each organism across the ditch was observed and the extent of inhibition from the edge of the ditch was determined. The extreme ends of the bacterial stripes served as controls to demonstrate the normal growth of the organism. As is the case with all antibacterial testing, it must be borne in mind that results may be subject to differing susceptibilities of different strains of an organism to the test material. marcescens
RESULTS
The antibacterial activities exhibited by the metal hydroxides per se are shown in Table 1. Testing of lathumycin in free form showed that it is active only against three out of the four organisms utilized, namely, E. coli, S. aureus, and S. faecalis. The results of the antibacterial testing of the insoluble derivatives of lathumycin are shown in Table 2, and those of the insoluble derivatives of other antibiotics are shown in Table 3. Estimates of the amounts of lathumycin ininobilized are also shown in Table 2. It was very difficult to obtain a zirconium hydroxide-agar mixture that would set to provide a suitable surface on which bacterial growth could be observed and distinguished. A satisfactory surface was obtained by mixing a higher proportion of agar with the sample. An even more conclusive result was achieved by use of naturally colored bacteria, namely, S. marcescens and C. violaceum to observe the extent of growth after the incubation period.
DISCUSSION Subsequent to the original discoveries that the hydroxides (hydrous oxides) of transition metals are suitable as matrixes for the immobilization of enzymes, etc. (1, 7), it has been found that they possess pseudobiological activities. They can act as an artificial enzyme, simulat-
TABLE 1. Antibacterial activities of metal hydroxides Growth of organismb Metal hydroxide"
.mreE. coli
Zr (IV) Ti (IV) Ti (III) Fe (III)
P.
aeruginosa
S.
aureus
S. marces-
S. viola-
cens
ceum
+++
+++
+++
+++
+++
+++
+++ (4)
+++ (2)
+++ (6) + +++ (2) +++
+++
-
-
.voa
S. faecalis
+++ (6) +++ +++ V(III) +++ +++ Sn(II)+ a All the hydroxides were prepared at pH 7.0. b Symbols: -, No inhibition of growth; +, slight inhibition of growth across ditch; + ++, complete inhibition of growth across ditch. Numbers in parentheses refer to maximal distance (millimeters) of
inhibition of growth of organism from the ditch.
768
ANTIMICROB. AGENTS CHEMOTHER.
KENNEDY AND HUMPHREYS
TABLE 2. Antibacterial activity of lathumycin coupled to insoluble metal hydroxides Growth of organisme
of antiMetal hydrox. ~~Amt hydrox Metal pH of coupling biotic coupled idea (mg)
E. col
nosa
S. aureus
S. faecalis
+++ (5) +++ 7.1 4.0 Zr(IV) +++ (7) +++ Zr(IV) 5.1 2.5 + 0.4 + Ti (IV) 7.1 +++ (7) +++ Ti (IV) 5.1 0.7 + 7.0 0.7 +++ +++ Ti (III) Ti (III) 5.1 0.6 +++ (3) +++ (7) 7.1 2.0 Fe (III) +++ +++ (5) 0.0 Fe (III) 5.1 + + +++ (5) +++ (5) V (III) 7.1 2.4 V (III) +++ 2.6 ++ 4.9 +++ (8) +++ (7) 3.8 7.1 Sn (II) _ +++ (5) 3.9 Sn (H) 5.0 +++ (7) a The parent antibiotic in free form was inactive against P. aeruginosa but was active against E. coli, S. aureus, and S. faecalis, inhibition occurring across the ditch and up to 25 mm from the ditch edge. b Symbols: -, No inhibition of growth; +, slight inhibition of growth across ditch; + + +, complete inhibition of growth across ditch. Numbers in parentheses refer to maximal distance (millimeters) of inhibition of growth of organisms from the ditch.
TABLE 3. Antibacterial activities of antibiotics insolubilized on zirconium hydroxide Growth of organisme Antibiotica S. au- S. faecalis E. coli nosa +++ +++ +++ Neomycin +++ +++ +++ Polymyxin B +++ (2) +++ (7) +++ Streptomycin +++ (8) _ +++ (5) + ++ (5) Chloramphen- +++ (11) _ icol +++ +++ Ampicillin + +++ + Penicillin G aAll parent antibiotics in free form were strongly active against all four organisms, inhibition occurring across the ditch and up to 25 mm from the ditch edge, with the exception of ampicillin, which was inactive against P. aerugi-
nosa.
bSymbols: -, No inhibition of growth; +, slight inhibition of growth across ditch; + + +, complete inhibition of growth across ditch. Numbers in parentheses refer to maximal distance (millimeters) of inhibition of growth of organism from the ditch.
ing the action of alkaline phosphatase (3), and the present work demonstrates that such hydroxides possess antibacterial activities per se. The results (Table 1) show that the antibacterial activities were exhibited largely by metals in lower oxidation states. It may well be that the antibacterial activity arises in some instances from the availability of electrons in the oxidation process, to which these compounds are susceptible. The lower oxidation states of these metal ions are unstable, as demonstrated by the color changes which occur upon exposure to the atmosphere. No doubt such redox reactions, occurring in the presence of bacteria, can have a deleterious effect, resulting in inhibition of bacterial growth. Although we have demon-
strated (2) that cells can be immobilized on metal hydroxides, it is unlikely that the chelation of the hydroxide with the cell wall material in the present work is responsible for growth prevention, since the immobilized cells were shown (2) to continue their respiration and growth processes. Such a phenomenon indicates that the surfaces of a number of metal materials may indeed be autosterile. The difficulty experienced in obtaining a zirconium hydroxide-agar mixture that would provide a suitable surface against which to observe bacterial growth is attributable to the strong chelating ability of zirconium hydroxide. The hydroxyl groups in the agar are chelated by the hydroxide, resulting in destabilization of the gel, and the problem was somewhat overcome by elevation of the agar proportion. It was shown that the question of growth of the organism across the ditch in difficult circumstances could be clarified without alteration of the agar proportion. This could be achieved using a colored organism, S. marcescens, which when grown at 25 C (but not 37 C) produces a distinctive red coloration. Whereas the color in liquid culture does become partially extracellular, under our conditions in solidphase culture the color remains associated with the cells. This enables growth of the cells to be readily determined against the agar suspension background. The use of C. violaceum, which is colored purple, is analogous. The organism produces a purple color during growth at both 25 and 37 C, and under the conditions of the solidphase culture the color remains associated with the organism. Further advantages of use of this technique are that, if necessary, the organisms
VOL. 9, 1976
METAL HYDROXIDES AND IMMOBILIZED ANTIBIOTICS
can be distinguished from others also present and sterile conditions can be avoided. The mechanism of antibiotic chelation by the metal hydroxides is analogous to their chelation by cellulose-metal chelates, the mechanism for which has been discussed in a previous paper (4). In brief, it is considered that antibiotic attachment occurs by replacement of a hydroxyl ligand on the metal hydroxide surface with a suitable functional group of the antibiotic; in such a way the antibiotic acts as a ligand. Such functional groups include hydroxyl, carboxyl, amino, etc., i.e., groups having electrons available for donation to the partially vacant d orbitals of the metals, resulting in the formation of a partial covalent bond (Fig. 1). A more detailed treatment of the mechanism of attachment is not appropriate to the location of this paper, but further chemical aspects will be published (2). The strength of this bond will vary, depending on the ligand(s) involved, and will thus determine the stability of the antibiotic-metal hydroxide complex. The loading of lathumycin on the metal hydroxides (Table 2) which was achieved does not appear to be a direct function of the pH of the preparation of hydroxide, each metal and pH exerting its own independent properties. However, it is clear that lathumycin derivatives of various metal hydroxides can be prepared and that they possess antibacterial activity (Table 2). More detailed interpretation of these data must be subject to a number of phenomena, as follows. The activities observed may be a compounding of antibiotic antibacterial activity and metal hydroxide antibacterial
activity. Where the metal hydroxide per se does not exhibit such activity (Table 1) it may be assumed that the activity of the antibiotic-hydroxide chelate is due to the antibiotic. Where the metal hydroxide per se does exhibit antibacterial activity a number of situations were found (Table 2) as follows. In the case where growth of P. aeruginosa, which is not susceptible to lathumycin, was inhibited and lathumycin-susceptible organisms were inhibited by the complex, the inhibition of P. aeruginosa is due to the action of the metal hydroxide, whereas the inhibition of the other three organisms can be due to antibiotic or hydroxide or both. In the cases where growth of P. aeruginosa was not inhibited by the complex, yet lathumycin-susceptible organisms were inhibited by the complex, it may be concluded that this inhibition is due to antibiotic rather than to metal hydroxide and that the presence of the antibiotic actually blocks the
769
0
R
,~r O HO/ >4QOt~7 V Zr
Zr o
0
0
0
FIG. 1. Projected structures of the complexes of zirconium hydroxide with compounds bearing carboxyl, hydroxyl, and amino groups, respectively.
action by the hydroxide. In the cases where the complex did not manifest any activity against a lathumycin-susceptible organism, it is clear that the formation of the complex has blocked both metal hydroxide activity and antibiotic activity so far as that organism is concerned. The case of immobilization on titanium (III) hydroxide is unique in that the uncomplexed hydroxide was active in every case, whereas the complexed hydroxide (pH 5.0), which had certainly retained some antibiotic, was inactive in every case; this presents the only case in the lathumycin series in which chelation blocked
770 KENNEDY AND HUMPHREYS both hydroxide and antibiotic activity throughout. Such may be due to the high reducing power of the titanium (III) ion acting upon the antibiotic. Immobilization of a wider range of antibiotics on zirconium (IV) hydroxide gave products that showed (Table 3) spectra of activities analogous to those found for such antibiotics immobilized on polymeric organic matrixes (4-6). Again the blocking effect of the bound antibiotic upon the activity of the metal hydroxide was apparent, thereby demonstrating that the activities recognized were due to the antibiotics concerned. As in previous work (4-6), the manifestation of antibacterial activity away from the ditch is considered to be due to slow release and diffusion of the antibiotic. Slow-release properties in the case of immobilization by chelation with metal hydroxides are due to different stabilities of the antibiotic-metal hydroxide described above. However, two other phenomena, not previously applicable, may also have aided diffusion. First, since the hydroxides are small molecules, their chelation with agar may tend to solubilize the agar gel and therefore provide a means of transport. Second, since the antibiotics are originally held by chelation, it may be that certain functional groups arising from components of the agar medium compete with the antibiotic for ligand sites and, where the antibiotic-metal bond is weak, cause a displacement. In conclusion, this work demonstrates that certain transition metal hydroxides possess antibacterial activity and that it is possible to form active, immobilized antibiotics by a simple process of chelation with metal hydroxides at around neutral pH in aqueous solution. The simplicity, ease of use, and cheapness of the reagent recommend this technique as a means of antibiotic immobilization, and by judicious
ANTIMICROB. AGENTS CHEMOTHER.
choice of metal hydroxide a range of antibacterial effects may be obtained. A particular advantage of the immobilization method is that, unlike all other methods, it does not require the preparation of the matrix, since this can be prepared in situ and on site using simple stable reagents without the aid of sophisticated equipment. The success of the preparation and activity of the matrix is guaranteed and, unlike the majority of matrixes, does not require use of spectroscopy, titration, etc. to confirm the reactivity of the preparation in hand. ACKNOWLEDGMENTS We thank Gist-Brocades, N. V., for provision of a research grant, a studentship (to J.D.H.), and a gift of lathumycin.
LITERATURE CITED 1. Kennedy, J. F., S. A. Barker, and J. D. Humphreys. 1976. Insoluble complexes of amino acids, peptides and enzymes with metal hydroxides. J. Chem. Soc. Perkin Trans. 1, in press. 2. Kennedy, J. F., S. A. Barker, and J. D. Humphreys. 1976. The immobilization of microbial cells on some metal hydroxides. Nature (London), in press. 3. Kennedy, J. F., S. A. Barker, and J. D. Humphreys. 1976. The zirconium hydroxide mediated, alkaline hydrolysis of phosphate esters. J. Chem. Soc. Perkin Trans. 1, in press. 4. Kennedy, J. F., S. A. Barker, and A. Zamir. 1974. Active insolubilized antibiotics based on cellulosemetal chelates. Antimicrob. Agents Chemother. 6:777-782. 5. Kennedy, J. F., and H. Cho Tun. 1973. Active insolubilized antibiotics based on cellulose and cellulose carbonate. Antimicrob. Agents Chemother. 3:575-579. 6. Kennedy, J. F., J. Epton, and G. R. Kennedy. 1973. Poly(N-acryloyl-4- and -5-aminosalicylic acids). II. Antibacterial properties and uses for the preparation of active, insoluble antibiotics. Antimicrob. Agents Chemother. 3:29-32. 7. Kennedy, J. F., and I. M. Kay. 1976. Hydrous titanium oxides-new supports for the simple immobilization of enzymes. J. Chem. Soc. Perkin Trans. 1, p. 329335.
p.
766-770
Vol. 9, No. 5 Printed in U.SA.
Active Immobilized Antibiotics Based on Metal Hydroxides1 JOHN F. KENNEDY* AND JOHN D. HUMPHREYS
Department of Chemistry, University of Birmingham, Edgbaston, Birmingham B15 2TT, England Received for publication 22 December 1975
The water-insoluble hydroxides of zirconium (IV), titanium (IV), titanium (III), iron (II), vanadium (Ill), and tin (II) have been used to prepare insoluble derivatives of a cyclic peptide antibiotic by a facile chelation process. Testing of the antibacterial activities of the products against two gram-positive and two gram-negative bacteria showed that in the majority of cases the water-insoluble antibiotics remained active against those bacteria susceptible to the parent antibiotic. The power of the assay system has been extended by the novel use of colored organisms to aid determinations where the growth of normal organisms could not be distinguished from the appearance of the supporting material. Insoluble derivatives of neomycin, polymyxin B, streptomycin, ampicillin, penicillin G, and chloramphenicol were prepared by chelation with zirconium hydroxide, and these derivatives similarly reflected the antibacterial activities of the parent compounds. Several of the metal hydroxides themselves possess antibacterial activity due to complex formation with the bacteria. However, the use of selected metal hydroxides can afford a simple, inexpensive, and inert matrix for antibiotic immobilization, resulting in an antibacterial product that may possess slow-release properties. The mechanisms by which the metal hydroxide-antibiotic association-dissociation may occur are discussed.
The potential uses of active, insolubilized antibiotic derivatives for the protection of biodegradable substances from microbial attack and for slow-release formulations in medical and industrial fields have already been discussed (4-6). Matrixes used for the formation of such derivatives have included cellulose (5), cellulose carbonate (5), poly(N-acryloyl-4- and -5aminosalicylic acids) (6) and cellulose-metal chelates (4). Whereas the formation of a completely covalent bond between the antibiotic and the chosen insoluble matrix provides the most stable immobilized antibiotic system, such a reaction may not always be feasible in practice and complete stability may not be desirable (e.g., in slow-release situations). Thus, the use of a simpler means of attachment can confer advantages of ease of preparation and also of economy. This has been demonstrated by the use of cellulose-metal chelates to insolubilize a range of antibiotics with retention of antibacterial activity (4). However, even in these cases, prior preparation of the activated matrix for immobilization is necessary. We now report the immobilization of antibiotics by a simple coupling process to matrixes which may be prepared as required without any need for specialized equipment. The matrixes are the hydroxides (hydrous oxides, hydroxyoxides) of the metal components of the cellulose-metalI
U.S. Patent 3,912,593.
antibiotic chelates (4). Such matrixes have already been used for the insolubilization of enzymes (1, 7), and the extension ofthis technique to the formation of insolubilized antibiotics has been studied and is now reported. MATERIALS AND METHODS Antibiotics. Antibiotics were obtained as follows: lathumycin, Gist-Brocades N.V.; neomycin sulfate, Boots Pure Drug Co. Ltd.; streptomycin sulfate, Glaxo Laboratories Ltd.; polymyxin B sulfate, Burroughs-Wellcome & Co.; ampicillin, Beecham Research Laboratories; chloramphenicol, Aspro-Nicholas Ltd.; penicillin G sodium salt, Imperial Chemical Industries, Ltd., Pharmaceuticals Division. Bacteria. The sources of bacteria used were: Escherichia coli NCTC 86, Pseudomonas aeruginosa BUCD 53 (Department of Chemistry, University of Birmingham), Streptococcus faecalis NCTC 370, Staphylococcus aureus (formerly S. pyogenes) NCTC 7447, Serratis marcescens BUCD 30, and Chromobacterium violaceum NCIB 9131. Metal chlorides. Iron (III) chloride (Hopkin and Williams Ltd.), tin (II) chloride (BDH Chemicals Ltd.), vanadium (III) chloride (BDH Chemicals Ltd.), and zirconium (IV) chloride (BDH Chemicals Ltd.) were standard laboratory reagents. Titanium (III) chloride (BDH Chemicals Ltd.) was supplied as a 12.5% (wt/vol) solution in hydrochloric acid and titanium (IV) chloride (BDH Chemicals Ltd.) was supplied as a 15% (wt/vol) solution in hydrochloric acid (5 N). Preparation of metal hydroxides. A standard method was used for the preparation of the various 766
VOL. 9, 1976
METAL HYDROXIDES AND IMMOBILIZED ANTIBIOTICS
metal hydroxides from solutions of the appropriate metal chloride above. The solutions of the metal chlorides were prepared by dissolving the requisite amount of the metal chloride in hydrochloric acid (1.0 M), except for the solution of tin (II) chloride, which was prepared by dissolving the chloride in 5.0 M hydrochloric acid to give final concentrations of 0.65 M. Titanium chloride solutions were used as supplied. The metal hydroxides were precipitated by the slow addition of ammonium hydroxide (2 M) to the metal chloride solutions (1.3 mmol of metal) to give the desired pH and to ensure that the gelatinous suspension formed was homogenous. Coupling of antibiotics to metal hydroxides. Two suspensions of the hydroxides of each of the metals listed above were prepared as described above, one suspension being adjusted to pH 5 and the other to pH 7. To every sample was added a 0. 1% (wt/vol) solution (5.0 ml) of lathumycin, a cyclic peptide antibiotic, and distilled water to give a final volume of 10.0 ml. After stirring for 2 h at 22 C, the mixtures were centrifuged and the absorbances of the supernatants were measured (to determine the amount of antibiotic remaining in solution). After washing with water (3 x 5.0 ml) the samples were tested for antibacterial activity as described below. A further six antibiotics were coupled to zirconium hydroxide at pH 7 in a similar manner to that described above, and the metal hydroxides (pH 7.0) were also prepared without any added antibiotic, to act as controls. Antibacterial testing of immobilized antibiotics. Ditch plates were prepared by allowing nutrient agar to solidify in a petri dish and a strip of agar, approximately 1 cm wide, was removed to form the ditch. The samples of immobilized antibiotics were resuspended in water (total volume, 10 ml), and a portion of the suspension was mixed with an equal volume of double-strength agar (1.6%, wt/vol) and introduced into the ditch. The plates were inoculated (in stripes perpendicular to the ditch) with two gram-negative (E. coli and P. aeruginosa) and two gram-positive (S. faecalis and S. aureus) bacteria, four stripes in all. In one case, two additional gramnegative bacteria were used (S. marcescens and C. violaceum) for which a second plate was used, two stripes in all.
767
The plates were incubated at 37 C (25 C for S. and C. violaceum) for 18 to 24 h, and then the degree of inhibition of growth of each organism across the ditch was observed and the extent of inhibition from the edge of the ditch was determined. The extreme ends of the bacterial stripes served as controls to demonstrate the normal growth of the organism. As is the case with all antibacterial testing, it must be borne in mind that results may be subject to differing susceptibilities of different strains of an organism to the test material. marcescens
RESULTS
The antibacterial activities exhibited by the metal hydroxides per se are shown in Table 1. Testing of lathumycin in free form showed that it is active only against three out of the four organisms utilized, namely, E. coli, S. aureus, and S. faecalis. The results of the antibacterial testing of the insoluble derivatives of lathumycin are shown in Table 2, and those of the insoluble derivatives of other antibiotics are shown in Table 3. Estimates of the amounts of lathumycin ininobilized are also shown in Table 2. It was very difficult to obtain a zirconium hydroxide-agar mixture that would set to provide a suitable surface on which bacterial growth could be observed and distinguished. A satisfactory surface was obtained by mixing a higher proportion of agar with the sample. An even more conclusive result was achieved by use of naturally colored bacteria, namely, S. marcescens and C. violaceum to observe the extent of growth after the incubation period.
DISCUSSION Subsequent to the original discoveries that the hydroxides (hydrous oxides) of transition metals are suitable as matrixes for the immobilization of enzymes, etc. (1, 7), it has been found that they possess pseudobiological activities. They can act as an artificial enzyme, simulat-
TABLE 1. Antibacterial activities of metal hydroxides Growth of organismb Metal hydroxide"
.mreE. coli
Zr (IV) Ti (IV) Ti (III) Fe (III)
P.
aeruginosa
S.
aureus
S. marces-
S. viola-
cens
ceum
+++
+++
+++
+++
+++
+++
+++ (4)
+++ (2)
+++ (6) + +++ (2) +++
+++
-
-
.voa
S. faecalis
+++ (6) +++ +++ V(III) +++ +++ Sn(II)+ a All the hydroxides were prepared at pH 7.0. b Symbols: -, No inhibition of growth; +, slight inhibition of growth across ditch; + ++, complete inhibition of growth across ditch. Numbers in parentheses refer to maximal distance (millimeters) of
inhibition of growth of organism from the ditch.
768
ANTIMICROB. AGENTS CHEMOTHER.
KENNEDY AND HUMPHREYS
TABLE 2. Antibacterial activity of lathumycin coupled to insoluble metal hydroxides Growth of organisme
of antiMetal hydrox. ~~Amt hydrox Metal pH of coupling biotic coupled idea (mg)
E. col
nosa
S. aureus
S. faecalis
+++ (5) +++ 7.1 4.0 Zr(IV) +++ (7) +++ Zr(IV) 5.1 2.5 + 0.4 + Ti (IV) 7.1 +++ (7) +++ Ti (IV) 5.1 0.7 + 7.0 0.7 +++ +++ Ti (III) Ti (III) 5.1 0.6 +++ (3) +++ (7) 7.1 2.0 Fe (III) +++ +++ (5) 0.0 Fe (III) 5.1 + + +++ (5) +++ (5) V (III) 7.1 2.4 V (III) +++ 2.6 ++ 4.9 +++ (8) +++ (7) 3.8 7.1 Sn (II) _ +++ (5) 3.9 Sn (H) 5.0 +++ (7) a The parent antibiotic in free form was inactive against P. aeruginosa but was active against E. coli, S. aureus, and S. faecalis, inhibition occurring across the ditch and up to 25 mm from the ditch edge. b Symbols: -, No inhibition of growth; +, slight inhibition of growth across ditch; + + +, complete inhibition of growth across ditch. Numbers in parentheses refer to maximal distance (millimeters) of inhibition of growth of organisms from the ditch.
TABLE 3. Antibacterial activities of antibiotics insolubilized on zirconium hydroxide Growth of organisme Antibiotica S. au- S. faecalis E. coli nosa +++ +++ +++ Neomycin +++ +++ +++ Polymyxin B +++ (2) +++ (7) +++ Streptomycin +++ (8) _ +++ (5) + ++ (5) Chloramphen- +++ (11) _ icol +++ +++ Ampicillin + +++ + Penicillin G aAll parent antibiotics in free form were strongly active against all four organisms, inhibition occurring across the ditch and up to 25 mm from the ditch edge, with the exception of ampicillin, which was inactive against P. aerugi-
nosa.
bSymbols: -, No inhibition of growth; +, slight inhibition of growth across ditch; + + +, complete inhibition of growth across ditch. Numbers in parentheses refer to maximal distance (millimeters) of inhibition of growth of organism from the ditch.
ing the action of alkaline phosphatase (3), and the present work demonstrates that such hydroxides possess antibacterial activities per se. The results (Table 1) show that the antibacterial activities were exhibited largely by metals in lower oxidation states. It may well be that the antibacterial activity arises in some instances from the availability of electrons in the oxidation process, to which these compounds are susceptible. The lower oxidation states of these metal ions are unstable, as demonstrated by the color changes which occur upon exposure to the atmosphere. No doubt such redox reactions, occurring in the presence of bacteria, can have a deleterious effect, resulting in inhibition of bacterial growth. Although we have demon-
strated (2) that cells can be immobilized on metal hydroxides, it is unlikely that the chelation of the hydroxide with the cell wall material in the present work is responsible for growth prevention, since the immobilized cells were shown (2) to continue their respiration and growth processes. Such a phenomenon indicates that the surfaces of a number of metal materials may indeed be autosterile. The difficulty experienced in obtaining a zirconium hydroxide-agar mixture that would provide a suitable surface against which to observe bacterial growth is attributable to the strong chelating ability of zirconium hydroxide. The hydroxyl groups in the agar are chelated by the hydroxide, resulting in destabilization of the gel, and the problem was somewhat overcome by elevation of the agar proportion. It was shown that the question of growth of the organism across the ditch in difficult circumstances could be clarified without alteration of the agar proportion. This could be achieved using a colored organism, S. marcescens, which when grown at 25 C (but not 37 C) produces a distinctive red coloration. Whereas the color in liquid culture does become partially extracellular, under our conditions in solidphase culture the color remains associated with the cells. This enables growth of the cells to be readily determined against the agar suspension background. The use of C. violaceum, which is colored purple, is analogous. The organism produces a purple color during growth at both 25 and 37 C, and under the conditions of the solidphase culture the color remains associated with the organism. Further advantages of use of this technique are that, if necessary, the organisms
VOL. 9, 1976
METAL HYDROXIDES AND IMMOBILIZED ANTIBIOTICS
can be distinguished from others also present and sterile conditions can be avoided. The mechanism of antibiotic chelation by the metal hydroxides is analogous to their chelation by cellulose-metal chelates, the mechanism for which has been discussed in a previous paper (4). In brief, it is considered that antibiotic attachment occurs by replacement of a hydroxyl ligand on the metal hydroxide surface with a suitable functional group of the antibiotic; in such a way the antibiotic acts as a ligand. Such functional groups include hydroxyl, carboxyl, amino, etc., i.e., groups having electrons available for donation to the partially vacant d orbitals of the metals, resulting in the formation of a partial covalent bond (Fig. 1). A more detailed treatment of the mechanism of attachment is not appropriate to the location of this paper, but further chemical aspects will be published (2). The strength of this bond will vary, depending on the ligand(s) involved, and will thus determine the stability of the antibiotic-metal hydroxide complex. The loading of lathumycin on the metal hydroxides (Table 2) which was achieved does not appear to be a direct function of the pH of the preparation of hydroxide, each metal and pH exerting its own independent properties. However, it is clear that lathumycin derivatives of various metal hydroxides can be prepared and that they possess antibacterial activity (Table 2). More detailed interpretation of these data must be subject to a number of phenomena, as follows. The activities observed may be a compounding of antibiotic antibacterial activity and metal hydroxide antibacterial
activity. Where the metal hydroxide per se does not exhibit such activity (Table 1) it may be assumed that the activity of the antibiotic-hydroxide chelate is due to the antibiotic. Where the metal hydroxide per se does exhibit antibacterial activity a number of situations were found (Table 2) as follows. In the case where growth of P. aeruginosa, which is not susceptible to lathumycin, was inhibited and lathumycin-susceptible organisms were inhibited by the complex, the inhibition of P. aeruginosa is due to the action of the metal hydroxide, whereas the inhibition of the other three organisms can be due to antibiotic or hydroxide or both. In the cases where growth of P. aeruginosa was not inhibited by the complex, yet lathumycin-susceptible organisms were inhibited by the complex, it may be concluded that this inhibition is due to antibiotic rather than to metal hydroxide and that the presence of the antibiotic actually blocks the
769
0
R
,~r O HO/ >4QOt~7 V Zr
Zr o
0
0
0
FIG. 1. Projected structures of the complexes of zirconium hydroxide with compounds bearing carboxyl, hydroxyl, and amino groups, respectively.
action by the hydroxide. In the cases where the complex did not manifest any activity against a lathumycin-susceptible organism, it is clear that the formation of the complex has blocked both metal hydroxide activity and antibiotic activity so far as that organism is concerned. The case of immobilization on titanium (III) hydroxide is unique in that the uncomplexed hydroxide was active in every case, whereas the complexed hydroxide (pH 5.0), which had certainly retained some antibiotic, was inactive in every case; this presents the only case in the lathumycin series in which chelation blocked
770 KENNEDY AND HUMPHREYS both hydroxide and antibiotic activity throughout. Such may be due to the high reducing power of the titanium (III) ion acting upon the antibiotic. Immobilization of a wider range of antibiotics on zirconium (IV) hydroxide gave products that showed (Table 3) spectra of activities analogous to those found for such antibiotics immobilized on polymeric organic matrixes (4-6). Again the blocking effect of the bound antibiotic upon the activity of the metal hydroxide was apparent, thereby demonstrating that the activities recognized were due to the antibiotics concerned. As in previous work (4-6), the manifestation of antibacterial activity away from the ditch is considered to be due to slow release and diffusion of the antibiotic. Slow-release properties in the case of immobilization by chelation with metal hydroxides are due to different stabilities of the antibiotic-metal hydroxide described above. However, two other phenomena, not previously applicable, may also have aided diffusion. First, since the hydroxides are small molecules, their chelation with agar may tend to solubilize the agar gel and therefore provide a means of transport. Second, since the antibiotics are originally held by chelation, it may be that certain functional groups arising from components of the agar medium compete with the antibiotic for ligand sites and, where the antibiotic-metal bond is weak, cause a displacement. In conclusion, this work demonstrates that certain transition metal hydroxides possess antibacterial activity and that it is possible to form active, immobilized antibiotics by a simple process of chelation with metal hydroxides at around neutral pH in aqueous solution. The simplicity, ease of use, and cheapness of the reagent recommend this technique as a means of antibiotic immobilization, and by judicious
ANTIMICROB. AGENTS CHEMOTHER.
choice of metal hydroxide a range of antibacterial effects may be obtained. A particular advantage of the immobilization method is that, unlike all other methods, it does not require the preparation of the matrix, since this can be prepared in situ and on site using simple stable reagents without the aid of sophisticated equipment. The success of the preparation and activity of the matrix is guaranteed and, unlike the majority of matrixes, does not require use of spectroscopy, titration, etc. to confirm the reactivity of the preparation in hand. ACKNOWLEDGMENTS We thank Gist-Brocades, N. V., for provision of a research grant, a studentship (to J.D.H.), and a gift of lathumycin.
LITERATURE CITED 1. Kennedy, J. F., S. A. Barker, and J. D. Humphreys. 1976. Insoluble complexes of amino acids, peptides and enzymes with metal hydroxides. J. Chem. Soc. Perkin Trans. 1, in press. 2. Kennedy, J. F., S. A. Barker, and J. D. Humphreys. 1976. The immobilization of microbial cells on some metal hydroxides. Nature (London), in press. 3. Kennedy, J. F., S. A. Barker, and J. D. Humphreys. 1976. The zirconium hydroxide mediated, alkaline hydrolysis of phosphate esters. J. Chem. Soc. Perkin Trans. 1, in press. 4. Kennedy, J. F., S. A. Barker, and A. Zamir. 1974. Active insolubilized antibiotics based on cellulosemetal chelates. Antimicrob. Agents Chemother. 6:777-782. 5. Kennedy, J. F., and H. Cho Tun. 1973. Active insolubilized antibiotics based on cellulose and cellulose carbonate. Antimicrob. Agents Chemother. 3:575-579. 6. Kennedy, J. F., J. Epton, and G. R. Kennedy. 1973. Poly(N-acryloyl-4- and -5-aminosalicylic acids). II. Antibacterial properties and uses for the preparation of active, insoluble antibiotics. Antimicrob. Agents Chemother. 3:29-32. 7. Kennedy, J. F., and I. M. Kay. 1976. Hydrous titanium oxides-new supports for the simple immobilization of enzymes. J. Chem. Soc. Perkin Trans. 1, p. 329335.
